Update app.py

app.py

@@ -2,72 +2,80 @@ import gradio as gr
 import joblib
 from huggingface_hub import hf_hub_download
 import numpy as np
-import pandas as pd  # Needed for DataFrame input to model
+import pandas as pd  # For DataFrame input to ensemble model
 
-# --- Download ensemble model from HF repo (single ensemble) ---
+# --- Define EnsembleModel class (same as Colab) ---
+class EnsembleModel:
+    def __init__(self, models):
+        self.models = models
+
+    def predict_proba(self, X):
+        # Average probabilities from all models
+        probs = [m.predict_proba(X)[:, 1] for m in self.models]
+        return np.mean(probs, axis=0)
+
+# --- Download ensemble from HF repo ---
 repo_id = "Ym420/terminator-ensemble-classification"
 ensemble_path = hf_hub_download(repo_id=repo_id, filename="ensemble.pkl")
-ensemble = joblib.load(ensemble_path)  # ✅ Load exactly as in Colab
+ensemble = joblib.load(ensemble_path)  # Load Colab ensemble
 
 # --- Bendability dictionary ---
 bend_dict = {
-    "AAA": -0.274, "AAC": -0.205, "AAG": -0.081, "AAT": -0.280,
-    "ACA": -0.006, "ACC": -0.032, "ACG": -0.033, "ACT": -0.183,
-    "AGA": 0.027, "AGC": 0.017, "AGG": -0.057, "AGT": -0.183,
-    "ATA": 0.182, "ATC": -0.110, "ATG": 0.134, "ATT": -0.280,
-    "CAA": 0.015, "CAC": 0.040, "CAG": 0.175, "CAT": 0.134,
-    "CCA": -0.246, "CCC": -0.012, "CCG": -0.136, "CCT": -0.057,
-    "CGA": -0.003, "CGC": -0.077, "CGG": -0.136, "CGT": -0.033,
-    "CTA": 0.090, "CTC": 0.031, "CTG": 0.175, "CTT": -0.081,
-    "GAA": -0.037, "GAC": -0.013, "GAG": 0.031, "GAT": -0.110,
-    "GCA": 0.076, "GCC": 0.107, "GCG": -0.077, "GCT": 0.017,
-    "GGA": 0.013, "GGC": 0.107, "GGG": -0.012, "GGT": -0.032,
-    "GTA": 0.025, "GTC": -0.013, "GTG": 0.040, "GTT": -0.205,
-    "TAA": 0.068, "TAC": 0.025, "TAG": 0.090, "TAT": 0.182,
-    "TCA": 0.194, "TCC": 0.013, "TCG": -0.003, "TCT": 0.027,
-    "TGA": 0.194, "TGC": 0.076, "TGG": -0.246, "TGT": -0.006,
-    "TTA": 0.068, "TTC": -0.037, "TTG": 0.015, "TTT": -0.274
+    "AAA": -0.274,"AAC": -0.205,"AAG": -0.081,"AAT": -0.280,
+    "ACA": -0.006,"ACC": -0.032,"ACG": -0.033,"ACT": -0.183,
+    "AGA": 0.027,"AGC": 0.017,"AGG": -0.057,"AGT": -0.183,
+    "ATA": 0.182,"ATC": -0.110,"ATG": 0.134,"ATT": -0.280,
+    "CAA": 0.015,"CAC": 0.040,"CAG": 0.175,"CAT": 0.134,
+    "CCA": -0.246,"CCC": -0.012,"CCG": -0.136,"CCT": -0.057,
+    "CGA": -0.003,"CGC": -0.077,"CGG": -0.136,"CGT": -0.033,
+    "CTA": 0.090,"CTC": 0.031,"CTG": 0.175,"CTT": -0.081,
+    "GAA": -0.037,"GAC": -0.013,"GAG": 0.031,"GAT": -0.110,
+    "GCA": 0.076,"GCC": 0.107,"GCG": -0.077,"GCT": 0.017,
+    "GGA": 0.013,"GGC": 0.107,"GGG": -0.012,"GGT": -0.032,
+    "GTA": 0.025,"GTC": -0.013,"GTG": 0.040,"GTT": -0.205,
+    "TAA": 0.068,"TAC": 0.025,"TAG": 0.090,"TAT": 0.182,
+    "TCA": 0.194,"TCC": 0.013,"TCG": -0.003,"TCT": 0.027,
+    "TGA": 0.194,"TGC": 0.076,"TGG": -0.246,"TGT": -0.006,
+    "TTA": 0.068,"TTC": -0.037,"TTG": 0.015,"TTT": -0.274
 }
 
-# --- Feature functions (match training exactly) ---
+# --- Feature functions (same as Colab) ---
 def gc_content(seq):
     seq = seq.upper()
     return (seq.count("G") + seq.count("C")) / len(seq) if len(seq) > 0 else 0
 
 def cpg_ratio(seq):
     seq = seq.upper()
-    length = len(seq)
-    if length == 0: return 0
+    l = len(seq)
+    if l == 0: return 0
     g = seq.count("G")
     c = seq.count("C")
     cg = seq.count("CG")
-    expected = (g * c) / length
+    expected = (g * c) / l
     return cg / expected if expected > 0 else 0
 
 def deltaG_stem_loop(seq):
     seq = seq.upper()
-    rna = seq.replace("T", "U")
-    nn = {
-        "AA": -0.9, "AU": -1.1, "UA": -1.3, "CA": -0.9,
-        "CU": -2.1, "GA": -1.3, "GU": -1.1, "UU": -0.9,
-        "AC": -0.9, "AG": -1.3, "UG": -1.5, "UC": -1.5,
-        "CC": -1.7, "CG": -2.4, "GC": -3.4, "GG": -1.5
-    }
+    rna = seq.replace("T","U")
+    nn = {"AA": -0.9,"AU": -1.1,"UA": -1.3,"CA": -0.9,
+          "CU": -2.1,"GA": -1.3,"GU": -1.1,"UU": -0.9,
+          "AC": -0.9,"AG": -1.3,"UG": -1.5,"UC": -1.5,
+          "CC": -1.7,"CG": -2.4,"GC": -3.4,"GG": -1.5}
     def rc(s):
-        comp = str.maketrans("ATCG", "TAGC")
+        comp = str.maketrans("ATCG","TAGC")
         return s.translate(comp)[::-1]
     deltaG = 0.0
     for i in range(len(seq)):
-        for j in range(i + 4, len(seq)):
+        for j in range(i+4,len(seq)):
             left = rna[i:j]
             right = rna[j:]
-            left_rc = rc(left).replace("T", "U")
+            left_rc = rc(left).replace("T","U")
             if left_rc in right:
                 total = 0.0
                 for k in range(len(left)-1):
                     pair = left[k:k+2]
                     if pair in nn: total += nn[pair]
-                if total < deltaG or deltaG == 0.0: deltaG = total
+                if total < deltaG or deltaG==0.0: deltaG = total
     return deltaG
 
 def avg_bendability(seq):
@@ -80,83 +88,61 @@ def avg_bendability(seq):
 
 def nucleotide_frequencies(seq):
     seq = seq.upper()
-    length = len(seq)
-    if length == 0: return 0,0,0,0
-    return seq.count("A")/length, seq.count("T")/length, seq.count("G")/length, seq.count("C")/length
+    l = len(seq)
+    if l == 0: return 0,0,0,0
+    return seq.count("A")/l, seq.count("T")/l, seq.count("G")/l, seq.count("C")/l
 
 def purine_pyrimidine_ratio(seq):
     seq = seq.upper()
     pur = seq.count("A")+seq.count("G")
     pyr = seq.count("C")+seq.count("T")
-    return pur/pyr if pyr > 0 else 0
+    return pur/pyr if pyr>0 else 0
 
-# --- Feature extraction ---
+# --- Extract features ---
 def extract_features(seq):
     gc = gc_content(seq)
     cpg = cpg_ratio(seq)
     dg = deltaG_stem_loop(seq)
     bend = avg_bendability(seq)
-    freq_a, freq_t, freq_g, freq_c = nucleotide_frequencies(seq)
+    freq_a,freq_t,freq_g,freq_c = nucleotide_frequencies(seq)
     pur_pyr = purine_pyrimidine_ratio(seq)
-    # ✅ Use SAME order as training
-    return [gc, cpg, dg, bend, freq_a, freq_t, freq_g, freq_c, pur_pyr]
+    return [gc, cpg, dg, bend, freq_a,freq_t,freq_g,freq_c, pur_pyr]
 
 # --- Prediction functions ---
 def predict_terminator(sequence: str) -> tuple[str, float]:
     clean_seq = "".join(sequence.split()).upper()
     X_new_df = pd.DataFrame([extract_features(clean_seq)], columns=[
-        "gc_content",
-        "cpg_ratio",
-        "deltaG",
-        "bendability",
-        "freq_A",
-        "freq_T",
-        "freq_G",
-        "freq_C",
-        "purine_pyrimidine_ratio"
+        "gc_content", "cpg_ratio", "deltaG", "bendability",
+        "freq_A","freq_T","freq_G","freq_C","purine_pyrimidine_ratio"
     ])
-    y_pred_proba = ensemble.predict_proba(X_new_df)[0]  # ✅ Single ensemble
-    label = "Terminator" if y_pred_proba >= 0.5 else "Non-terminator"
-    confidence = round(float(y_pred_proba), 4)
+    y_pred_proba = ensemble.predict_proba(X_new_df)[0]
+    label = "Terminator" if y_pred_proba>=0.5 else "Non-terminator"
+    confidence = round(float(y_pred_proba),4)
     return label, confidence
 
 def predict_terminator_table(sequence: str):
-    label, confidence = predict_terminator(sequence)
-    non_terminator_conf = round(1.0 - confidence, 4)
-    return [
-        ["Terminator", confidence],
-        ["Non-terminator", non_terminator_conf]
-    ]
+    label, conf = predict_terminator(sequence)
+    return [["Terminator", conf], ["Non-terminator", round(1-conf,4)]]
 
 # --- Gradio UI ---
-custom_css = """
-footer, .footer {
-    display: none !important;
-}
-"""
-
+custom_css = "footer, .footer {display:none !important;}"
 with gr.Blocks(css=custom_css, theme="default") as demo:
     gr.Markdown("## Intrinsic Terminator Prediction\nEnter a DNA sequence to predict terminator probability.")
-    
     seq = gr.Textbox(label="Enter DNA sequence")
-    
     with gr.Row():
         predict_btn = gr.Button("Predict", variant="primary", elem_id="predict-btn")
         clear_btn = gr.Button("Clear", elem_id="clear-btn")
-
     gr.HTML("""
         <style>
-        #predict-btn { width: 48%; min-width: 120px; }
-        #clear-btn { width: 48%; min-width: 100px; }
+        #predict-btn { width:48%; min-width:120px; }
+        #clear-btn { width:48%; min-width:100px; }
         </style>
-    """)   
-    
+    """)
     table = gr.Dataframe(headers=["Class","Confidence"], datatype=["str","number"], interactive=False)
-    
     predict_btn.click(fn=predict_terminator_table, inputs=seq, outputs=table)
-    clear_btn.click(fn=lambda: ("", []), outputs=[seq, table])
-    
+    clear_btn.click(fn=lambda: ("",[]), outputs=[seq, table])
     gr.api(predict_terminator, api_name="predict_terminator")
 
-if __name__ == "__main__":
+if __name__=="__main__":
     demo.launch()
+